The semiconductor industry continues to develop lithographic technologies which are able to print ever-smaller integrated circuit dimensions. Extreme ultraviolet (“EUV”) light (also sometimes referred to as soft x-rays) is generally defined to be electromagnetic radiation having wavelengths of between about 5 and 120 nm. EUV lithography is currently generally considered to include EUV light at wavelengths in the range of about 10-14 nm, and is used to produce extremely small features, for example, sub-32 nm features, in substrates such as silicon wafers. These systems must be highly reliable and provide cost effective throughput and reasonable process latitude.
Methods to produce EUV light include, but are not necessarily limited to, converting a material into a plasma state that has one or more elements, e.g., xenon, lithium, tin, indium, antimony, tellurium, aluminum, etc., with one or more emission line(s) in the EUV range. In one such method, often termed laser produced plasma (“LPP”), the required plasma can be produced by irradiating a target material, such as a droplet, stream or cluster of material having the desired line-emitting element, with a laser pulse at an irradiation site. The target material may contain the spectral line-emitting element in a pure form or alloy form, for example, an alloy that is a liquid at desired temperatures, or may be mixed or dispersed with another material such as a liquid.
In one common embodiment, a droplet generator heats the target material and extrudes the heated target material as a series of droplets that travel along a trajectory to the irradiation site to intersect a corresponding series of laser pulses. Ideally, the irradiation site is at one focal point of a reflective collector. When a laser pulse hits a droplet at the irradiation site, the droplet is vaporized and the reflective collector causes the resulting EUV light output to be maximized at another focal point of the collector. When subsequent droplets are hit with subsequent laser pulses, further EUV light output is provided.
LPP EUV systems are typically “MOPA” systems, in which a master oscillator and power amplifier form a source laser which may be fired as and when desired, and “MOPA PP” (“MOPA with pre-pulse”) systems in which a droplet is sequentially illuminated by more than one light pulse. In a MOPA PP system, a “pre-pulse” is first used to heat, vaporize or ionize the droplet and generate a weak plasma, followed by a “main pulse” which converts most or all of the droplet material into a strong plasma to produce EUV light emission.
One issue is that it is desirable, and in fact important, to be able to control the amount, or “dose,” of EUV light energy being applied to a particular item being treated, such as a semiconductor wafer. For example, a specified amount of EUV light energy may be required to accomplish some task, such as curing a layer of photoresist, on a semiconductor wafer as part of the manufacturing process. In order to obtain consistent results across different wafers, it will be desirable to apply the same amount of EUV light energy to each wafer, to as great a degree of accuracy as possible.
This is complicated by the fact that the power in each laser pulse may vary. Since the amount of EUV energy released when a laser pulse hits a droplet varies with the power in the laser pulse, the EUV light energy generated by any given droplet may also vary.
At present there are two ways that such dose control is accomplished in an EUV source. One is known as packet-based dose control, and the other is called EUV stabilization.
In packet-based dose control, the laser pulses, and thus the corresponding droplets, are divided into “packets” or groups of pulses (and droplets), and a desired dose target selected that each packet should meet. A packet may typically include 50 pulses, but packets of as few as 15 pulses have also been used. A dose target is selected, which each packet is intended to meet.
The EUV energy that is generated by each pulse hitting a corresponding droplet is measured. For each packet, a total accumulated dose is then calculated by adding the energy from each droplet over the series of droplets, starting with the first droplet in the packet. Once the dose target for the packet is achieved, the rest of the pulses in that packet are “skipped” or “missed,” i.e., droplets are not hit by the laser pulses. Skipping a droplet is typically accomplished either by firing the laser at a location other than the irradiation site at which the droplet is located, or by firing the laser at a time such that a droplet will not be at the irradiation site when the laser pulse arrives there.
The problem with packet-based dose control is that due to the variation in laser pulse energy, and thus in the EUV energy generated by each droplet, different packets may end up with very different numbers of pulses that actually generate energy. Since the pulses that do not generate EUV energy are all at the ends of the packets, there will be gaps between EUV pulse trains in sequential packets, and these gaps will also have a variable duration. In some cases, the target dose might be met by 10 droplets of a 15 droplet packet, with the remaining 5 droplets not hit, or 30 droplets of a 50 droplet packet, with the remaining 20 droplets not hit, resulting in gaps of 33% and 40% of the packet respectively.
The EUV energy also heats up the EUV plasma, and the variation in the EUV pulse trains will thus cause the temperature of the plasma to also vary from packet to packet. This variation in temperature can lead to a less stable plasma and in turn cause further variations in the EUV pulse energy. As a result, a larger “dose margin,” the difference between the maximum power that the system can theoretically produce and the amount of power that is desired, is required in order to insure that the dose target will be consistently met. This reduces the dose controlled EUV power that can be achieved in the LLP EUV source.
The EUV stabilization approach to dose control attempts to avoid plasma instabilities by eliminating the gaps between the packets that occur in packet-based dose control. Instead of skipping droplets, the pulse energy of each laser pulse is controlled by adjusting either the duration of the pulse or the magnitude of the pulse from the master oscillator of the laser.
The problem with the EUV stabilization approach is that reduction of the master oscillator pulse energy leads to incomplete extraction of the gain in the power amplifier. As is well known in the art, this can lead to self-lasing and large amounts of reflected power from the droplets that can damage optical components in the system. Additionally, reducing the laser energy hitting the droplet can produce increased amount of debris due to incomplete evaporation of the target, which is not desirable in the source.
What is needed is an improved way to control a dose of EUV radiation generated by an LPP EUV light source that extracts all of the gain from the power amplifier without creating large gaps between EUV pulse trains.